COF-C03 Question 2
Single answer1.1 Describe and use the Snowflake architectureA retail company uses Snowflake for both nightly ELT pipelines and daytime business intelligence dashboards. The data engineering team reports that increasing the warehouse size for the nightly ELT jobs improved load performance, but dashboard users still experienced slow response times during the day when many analysts queried the same tables. The architect wants to improve dashboard concurrency without affecting the ELT workload or requiring data copies. Which Snowflake architectural approach best addresses this requirement?
- A
Create a separate virtual warehouse for the BI dashboards so compute resources are isolated while both workloads access the same centrally stored data
- B
Duplicate the underlying tables into a second database dedicated to BI so dashboard queries do not compete with ELT operations
- C
Increase the maximum data retention period on the tables so more query versions can be served concurrently
- D
Move the BI users to the same larger warehouse used by ELT, because a single larger warehouse is the only way to improve concurrency in Snowflake
Show answer and explanation
Correct answer: A
Explanation
This question tests understanding of Snowflake's multi-cluster shared data architecture: centralized storage with independent compute in the form of virtual warehouses. A key architectural advantage is that different workloads, such as ELT and BI, can run on separate warehouses against the same underlying data without data duplication. This improves workload isolation, cost control, and performance management. In Snowflake documentation, virtual warehouses are described as independent compute resources that can be started, stopped, resized, and assigned to specific workloads. This is a foundational concept for SnowPro Core: storage and compute are decoupled, and concurrency or performance issues are often addressed by assigning workloads to separate warehouses rather than copying data or changing retention settings.
- A. Correct.
Correct. Snowflake's architecture separates storage from compute. Multiple virtual warehouses can access the same data in centralized storage without copying it. Creating a dedicated warehouse for BI isolates compute consumption and contention from ELT processing, which is a common best practice for mixed workloads requiring independent performance and scaling.
- B. Incorrect.
Incorrect. Duplicating tables is unnecessary for this scenario and works against one of Snowflake's core architectural benefits: shared centralized storage with independent compute layers. Separate warehouses, not copied data, are the standard approach for isolating workloads.
- C. Incorrect.
Incorrect. Data retention controls Time Travel behavior, not query concurrency or workload isolation. Increasing retention does not improve dashboard performance under concurrent demand and does not address compute contention between ELT and BI workloads.
- D. Incorrect.
Incorrect. A larger single warehouse may improve performance for some queries, but it does not provide workload isolation. ELT and BI would still share the same compute cluster, so resource contention could continue. Snowflake architecture supports separate warehouses specifically to isolate workloads while using the same data.